Skip to main content

Troubleshooting

For a full local or CI setup check, run:

wendell doctor --validate

The doctor command checks credentials, config loading, API identity, and agent_command parsing without printing secrets.

Not logged in​

Error:

Wendell error: no runner credential found.

Fix for a new local runner:

wendell register

Fix for CI:

export WENDELL_INKPASS_API_KEY=...

Fix when you already have an API key and want to store it locally:

printf "%s\n" "$WENDELL_INKPASS_API_KEY" | wendell login \
  --api-key-stdin \
  --validate

Required questions block approval​

Error:

Playbook Summary has unresolved required questions.

Fix:

wendell playbook review wdraft_abc123

Create a review patch with answer_question operations and apply it:

wendell playbook apply wdraft_abc123 \
  --file ./review.json

Then approve again.

Suite is not ready​

Error:

suite `...` is not bound to a ready scenario pack

Fix:

wendell playbook approve wdraft_abc123 --generate-suite
wendell suites publish --draft wdraft_abc123

Missing agent command​

Error:

wendell run requires `agent_command` in the config file

Fix:

Add an adapter command to wendell.toml:

agent_command = "python scripts/wendell_agent_adapter.py"

Runner is not authorized​

If a runner receives a 403 Forbidden, refresh credentials:

wendell register

Then retry:

wendell whoami
wendell suites configure --suite <suite-slug>
wendell run --suite <suite-slug>

If wendell run reports that wendell.toml is missing, run wendell suites configure --suite <suite-slug> from the application repo first. It creates the config and adapter template without embedding credentials.

Need machine-readable output​

Most commands support:

--json

Use JSON output for CI, automation, and agent workflows.